import pandas as pd
import numpy as np 
import matplotlib.pyplot as plt
import seaborn as sns

# Read data into a dataframe: 'merged_tourism.csv'

merged_tourism = pd.read_csv('merged_tourism.csv')

# Changed 'merged_tourism' columns names for 'tourism_share_gdp' to 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists' to 'Inbound Arrivals', 'business_personal_ratio' to 'Business-Personal Ratio'

merged_tourism.rename(columns={'tourism_share_gdp': 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists': 'Inbound Arrivals', 'business_personal_ratio': 'Business-Personal'}, inplace=True)
merged_tourism

merged_tourism['Entity'].unique()

array(['Albania', 'Antigua and Barbuda', 'Argentina', 'Aruba',
       'Azerbaijan', 'Bahamas', 'Bahrain', 'Belarus', 'Belgium',
       'Bermuda', 'Bhutan', 'Botswana', 'Brunei', 'Canada', 'Chile',
       'Colombia', 'Costa Rica', 'Denmark', 'Egypt', 'El Salvador',
       'Eswatini', 'Fiji', 'Finland', 'France', 'French Polynesia',
       'Guam', 'Guinea', 'Guyana', 'Honduras', 'Hong Kong', 'Hungary',
       'India', 'Indonesia', 'Ireland', 'Israel', 'Italy', 'Jamaica',
       'Jordan', 'Kazakhstan', 'Kiribati', 'Latvia', 'Lithuania',
       'Luxembourg', 'Malawi', 'Malaysia', 'Malta', 'Marshall Islands',
       'Martinique', 'Mauritius', 'Mexico', 'Micronesia (country)',
       'Mongolia', 'Montserrat', 'Morocco', 'Mozambique', 'Namibia',
       'Nepal', 'New Zealand', 'Nigeria', 'Norway', 'Oman', 'Palau',
       'Panama', 'Paraguay', 'Peru', 'Philippines', 'Poland', 'Reunion',
       'Rwanda', 'Saint Lucia', 'Saudi Arabia', 'Sierra Leone',
       'Singapore', 'Slovenia', 'South Africa', 'Spain', 'Sri Lanka',
       'Sweden', 'Thailand', 'Uganda', 'United Arab Emirates',
       'United Kingdom', 'United States', 'Uruguay', 'Vietnam'],
      dtype=object)

# Read data into a dataframe: 'merged_tourism_localvforeign.csv'

merged_tourismLF = pd.read_csv('merged_tourism_localvforeign.csv')

# Changed 'merged_tourismLF' columns names for 'tourism_share_gdp' to 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists' to 'Inbound Arrivals', 'business_personal_ratio' to 'Business-Personal'
# 'dom_tour_accom_hotels_and_similar_estab_guests' to 'Domestic tourists', 'in_tour_accom_hotels_and_similar_estab_guests' to 'Foreign guests'

merged_tourismLF.rename(columns={'tourism_share_gdp': 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists': 'Inbound Arrivals', 'business_personal_ratio': 'Business-Personal', 'dom_tour_accom_hotels_and_similar_estab_guests': 'Domestic tourists', 'in_tour_accom_hotels_and_similar_estab_guests': 'Foreign guests'}, inplace=True)
merged_tourismLF

merged_tourismLF['Entity'].unique()

array(['Albania', 'Argentina', 'Azerbaijan', 'Bahrain', 'Belarus',
       'Belgium', 'Chile', 'Denmark', 'Eswatini', 'Finland', 'France',
       'Hungary', 'Indonesia', 'Israel', 'Italy', 'Jordan', 'Latvia',
       'Lithuania', 'Luxembourg', 'Malaysia', 'Malta', 'Mexico',
       'Morocco', 'Mozambique', 'Norway', 'Peru', 'Poland',
       'Saudi Arabia', 'Slovenia', 'Spain', 'Sweden', 'Thailand',
       'United Arab Emirates', 'United Kingdom'], dtype=object)

# Read data into a dataframe: 'merged_tourism_inboundvoutbound.csv'

merged_tourismIO = pd.read_csv('merged_tourism_inboundvoutbound.csv')

# Changed 'merged_tourismLF' columns names for 'tourism_share_gdp' to 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists' to 'Inbound Arrivals', 'business_personal_ratio' to 'Business-Personal'
# 'inbound_outbound_tourism' to 'Inbound-Outbound'

merged_tourismIO.rename(columns={'tourism_share_gdp': 'Share GDP', 'in_tour_arrivals_ovn_vis_tourists': 'Inbound Arrivals', 'business_personal_ratio': 'Business-Personal', 'inbound_outbound_tourism': 'Inbound-Outbound'}, inplace=True)
merged_tourismIO

merged_tourismIO['Entity'].unique()

array(['Argentina', 'Azerbaijan', 'Bahrain', 'Belarus', 'Belgium',
       'Bermuda', 'Bhutan', 'Canada', 'Chile', 'Colombia', 'Costa Rica',
       'Denmark', 'El Salvador', 'Eswatini', 'Fiji', 'Finland', 'France',
       'French Polynesia', 'Honduras', 'Hungary', 'India', 'Israel',
       'Italy', 'Jordan', 'Kazakhstan', 'Latvia', 'Lithuania',
       'Luxembourg', 'Malawi', 'Malta', 'Mauritius', 'Mexico', 'Morocco',
       'Nepal', 'New Zealand', 'Norway', 'Oman', 'Panama', 'Paraguay',
       'Peru', 'Philippines', 'Poland', 'Saudi Arabia', 'Sierra Leone',
       'Singapore', 'Slovenia', 'Spain', 'Sri Lanka', 'Sweden',
       'Thailand', 'Uganda', 'United Arab Emirates', 'United States',
       'Uruguay'], dtype=object)

# merged_tourism shape
print(f"The merged_tourism dataset has {merged_tourism.shape[0]} rows and {merged_tourism.shape[1]} columns.")

# merged_tourismLF
print(f"The merged_tourismLF dataset has {merged_tourismLF.shape[0]} rows and {merged_tourismLF.shape[1]} columns.")

#merged_tourisoIO
print(f"The merged_tourismIO dataset has {merged_tourismIO.shape[0]} rows and {merged_tourismIO.shape[1]} columns.")

The merged_tourism dataset has 643 rows and 5 columns.
The merged_tourismLF dataset has 234 rows and 7 columns.
The merged_tourismIO dataset has 371 rows and 6 columns.

# merged_tourism description
merged_tourism.describe()

# merged_tourismLF description
merged_tourismLF.describe()

# merged_tourismIO description
merged_tourismIO.describe()

import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np

fig, axs = plt.subplots(1, 3, figsize=(24, 8))  # Increased figure size

# Heatmap/Correlation for merged_tourism
numerical_features = merged_tourism.select_dtypes(include=np.number)
sns.heatmap(numerical_features.corr(), annot=True, cmap='coolwarm', ax=axs[0], annot_kws={"size": 12})  # Increased font size
axs[0].set_title('Correlation Heatmap for merged_tourism', fontsize=14)  # Increased title font size

# Heatmap/Correlation for merged_tourismLF
numerical_featuresLF = merged_tourismLF.select_dtypes(include=np.number)
sns.heatmap(numerical_featuresLF.corr(), annot=True, cmap='coolwarm', ax=axs[1], annot_kws={"size": 12})
axs[1].set_title('Correlation Heatmap for merged_tourismLF', fontsize=14)

# Heatmap/Correlation for merged_tourismIO
numerical_featuresIO = merged_tourismIO.select_dtypes(include=np.number)
sns.heatmap(numerical_featuresIO.corr(), annot=True, cmap='coolwarm', ax=axs[2], annot_kws={"size": 12})
axs[2].set_title('Correlation Heatmap for merged_tourismIO', fontsize=14)

plt.tight_layout()
plt.subplots_adjust(wspace=0.3)  # Increased spacing between subplots
plt.show()

import plotly.express as px

# Creating a Bubble chart to explore 'Tourism Share of GDP vs. Year (Inbound Arrivals)
fig_mergedtourism = px.scatter(merged_tourism, 
                 x="Year", 
                 y="Share GDP", 
                 size="Inbound Arrivals", 
                 color="Entity",
                 hover_name="Entity", 
                 title="Tourism Share of GDP vs. Year (Bubble Size: Inbound Arrivals)",
                 labels={"Share GDP": "Tourism Share of GDP (%)", "Inbound Arrivals": "Inbound Tourist Arrivals"},
                 animation_frame="Year",
                 range_x=[merged_tourism['Year'].min() - 1, merged_tourism['Year'].max() + 1],
                 range_y=[merged_tourism['Share GDP'].min() - 1, merged_tourism['Share GDP'].max() + 1]
                )

fig_mergedtourism.update_layout(
    title_font=dict(size=20),
    xaxis_title_font=dict(size=16),
    yaxis_title_font=dict(size=16),
    legend_title_font=dict(size=16),
    legend=dict(itemsizing='constant') 
)

fig_mergedtourism.update_traces(marker=dict(sizeref=0.1))  # Make bubbles bigger

# To look at the visualization play the start button
fig_mergedtourism.show()

import plotly.express as px

# Calculate the ratio of domestic to foreign tourists
merged_tourismLF['Domestic/Foreign Ratio'] = merged_tourismLF['Domestic tourists'] / merged_tourismLF['Foreign guests']

# Creating a Bubble chart to explore 
fig_mergedtourismLF = px.scatter(merged_tourismLF,
                    x="Year",
                    y="Share GDP",
                    size="Domestic/Foreign Ratio",
                    color="Entity",
                    hover_name="Entity",
                    title="Tourism Share of GDP vs. Year (Bubble Size: Domestic/Foreign Tourists)",
                    labels={"Share GDP": "Tourism Share of GDP (%)", "Domestic/Foreign tourists": "Domestic-Foreign Tourist Arrivals"},
                    animation_frame="Year",
                    range_x=[merged_tourismLF['Year'].min() - 1, merged_tourismLF['Year'].max() + 1],
                    range_y=[merged_tourismLF['Share GDP'].min() - 1, merged_tourismLF['Share GDP'].max() + 1]
                   )

fig_mergedtourismLF.update_layout(
    title_font=dict(size=20),
    xaxis_title_font=dict(size=16),
    yaxis_title_font=dict(size=16),
    legend_title_font=dict(size=16),
    legend=dict(itemsizing='constant')
)

fig_mergedtourismLF.update_traces(marker=dict(sizeref=0.1))  # Make bubbles bigger

# To look at the visualization play the start button
fig_mergedtourismLF.show()

import plotly.express as px

# Creating a Bubble chart to explore 'Tourism Share of GDP vs. Year (Inbound-Outbound Ratio)'
fig_mergedtourismIO = px.scatter(merged_tourismIO, 
                 x="Year", 
                 y="Share GDP", 
                 size="Inbound-Outbound", 
                 color="Entity",
                 hover_name="Entity", 
                 title="Tourism Share of GDP vs. Year (Bubble Size: Inbound-Outbound Ratio)",
                 labels={"Share GDP": "Tourism Share of GDP (%)", "Inbound-Outbound": "Inbound-Outbound Ratio"},
                 animation_frame="Year",
                 range_x=[merged_tourismIO['Year'].min() - 1, merged_tourismIO['Year'].max() + 1],
                 range_y=[merged_tourismIO['Share GDP'].min() - 1, merged_tourismIO['Share GDP'].max() + 1]
                )

fig_mergedtourismIO.update_layout(
    title_font=dict(size=20),
    xaxis_title_font=dict(size=16),
    yaxis_title_font=dict(size=16),
    legend_title_font=dict(size=16),
    legend=dict(itemsizing='constant') 
)

fig_mergedtourismIO.update_traces(marker=dict(sizeref=0.1))  # Make bubbles bigger

# To look at the visualization play the start button
fig_mergedtourismIO.show()

# sns pairplot for merged_tourism
plt.figure(figsize=(8, 6))
sns.pairplot(merged_tourism)
plt.show()

<Figure size 800x600 with 0 Axes>

# sns pairplot for merged_tourismLF
plt.figure(figsize=(8, 6))
sns.pairplot(merged_tourismLF)
plt.show()

<Figure size 800x600 with 0 Axes>

# sns pairplot for merged_tourismIO
plt.figure(figsize=(8, 6))
sns.pairplot(merged_tourismIO)
plt.show()

<Figure size 800x600 with 0 Axes>

from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler
import pandas as pd

# Select relevant columns
columns_for_pca = ["Share GDP", "Inbound Arrivals", "Business-Personal"]
pca_data = merged_tourism[columns_for_pca]

# Standardize data (mean=0, variance=1)
scaler = StandardScaler()
pca_data_scaled = scaler.fit_transform(pca_data)

# Apply PCA (reduce to 2 components)
pca = PCA(n_components=2, random_state=7)
pca_result = pca.fit_transform(pca_data_scaled)

# Convert PCA output to DataFrame
pca_df = pd.DataFrame(pca_result, columns=["PC1", "PC2"])
pca_df["Country"] = merged_tourism["Entity"]
pca_df["Year"] = merged_tourism["Year"]

pca_df.head()

# Select relevant columns for merged_tourismLF
columns_for_pca_mergedtourismLF = ["Share GDP", "Domestic tourists", "Foreign guests"]
pca_mergedtourismLF = merged_tourismLF[columns_for_pca_mergedtourismLF]

# Standardize data (mean=0, variance=1)
scaler_LF = StandardScaler()
pca_mergedtourismLF_scaled = scaler_LF.fit_transform(pca_data)

# Apply PCA (reduce to 2 components)
pca_LF = PCA(n_components=2, random_state=7)
pca_LF_result = pca_LF.fit_transform(pca_data_scaled)

# Convert PCA output to DataFrame
LF = pd.DataFrame(pca_LF_result, columns=["PC1_LF", "PC2_LF"])
LF["Country"] = merged_tourismLF["Entity"]
LF["Year"] = merged_tourismLF["Year"]

LF.head()

# Select relevant columns for merged_tourismIO
columns_for_pca_mergedtourismIO = ["Share GDP", "Inbound-Outbound"]
columns_for_pca_mergedtourismIO = merged_tourismIO[columns_for_pca_mergedtourismIO]

# Standardize data (means=0, variance=1)
scaler_IO = StandardScaler()
pca_mergedtourismIO_scaled = scaler_IO.fit_transform(columns_for_pca_mergedtourismIO)

# Apply PCA
pca_IO = PCA(n_components=2, random_state=7)
pca_IO_result = pca_IO.fit_transform(pca_mergedtourismIO_scaled)

# Convert PCA output to DataFrame
IO = pd.DataFrame(pca_IO_result, columns=["PC1_IO", "PC2_IO"])
IO["Country"] = merged_tourismIO["Entity"]
IO["Year"] = merged_tourismIO["Year"]

IO.head()

import pandas as pd
from sklearn.preprocessing import StandardScaler
from sklearn.decomposition import PCA
import matplotlib.pyplot as plt
import seaborn as sns

# Create a figure with 2 subplots (1 row, 2 columns)
fig, axs = plt.subplots(1, 2, figsize=(10, 5))  # 1 row, 2 columns

# PCA projection plot for merged_tourism (Plot 1)
sns.scatterplot(x="PC1", y="PC2", hue="Country", data=pca_df, palette="viridis", alpha=0.7, ax=axs[0])
axs[0].set_xlabel("Principal Component 1", fontsize=12)
axs[0].set_ylabel("Principal Component 2", fontsize=12)
axs[0].set_title("PCA Projection of merged_tourism Data", fontsize=14)
axs[0].legend([], [])  # Hide long legend

# PCA projection plot for merged_tourismIO (Plot 2)
sns.scatterplot(x="PC1_IO", y="PC2_IO", hue="Country", data=IO, palette="magma", alpha=0.7, ax=axs[1])  # Changed ax to axs[1]
axs[1].set_xlabel("Principal Component 1", fontsize=12)
axs[1].set_ylabel("Principal Component 2", fontsize=12)
axs[1].set_title("PCA Projection of merged_tourismIO Data", fontsize=14)
axs[1].legend([], [], frameon=False)  # Hide long legend

# Adjust layout and spacing
plt.tight_layout()

# Display the figure
plt.show()

from sklearn.cluster import KMeans

# Elbow method for merged_tourism
X = pca_df[["PC1", "PC2"]]

# Try different k-values
clusters = [] # Initialize clusters as an empty list
K_range = range(1, 11)
for k in K_range:
    kmeans = KMeans(n_clusters=k, random_state=42, n_init=10)
    # Fit kmeans to the PCA data (X)
    kmeans.fit(X)
    clusters.append(kmeans.inertia_) # Append only inertia values to clusters

# Elbow method for merged_tourismIO
X_IO = IO[["PC1_IO", "PC2_IO"]]

# Try different k-values
clusters_IO = []  # Initialize clusters_IO as an empty list
K_range_IO = range(1, 11)
for k in K_range_IO:
    kmeans_IO = KMeans(n_clusters=k, random_state=42, n_init=10)
    # Fit kmeans to the PCA data (X)
    kmeans_IO.fit(X_IO) # Make sure to fit on X_IO, not X
    clusters_IO.append(kmeans_IO.inertia_)  # Append only inertia values to clusters_IO

from sklearn.cluster import KMeans
import matplotlib.pyplot as plt
import seaborn as sns

# Create a figure with 2 subplots (1 row, 2 columns)
fig, axs = plt.subplots(1, 2, figsize=(18, 6))  # Adjust figsize as needed

# Elbow method for merged_tourism (Plot 1)
axs[0].plot(K_range, clusters, marker='o', linestyle='-', color='b', linewidth=2)
axs[0].set_xlabel('Number of Clusters (k)', fontsize=12)
axs[0].set_ylabel('Within-Cluster Sum of Squares (WCSS)', fontsize=12)
axs[0].set_title('Elbow Method for merged_tourism', fontsize=14)
axs[0].grid(True, linestyle='--', alpha=0.7)

# Elbow method for merged_tourismIO (Plot 2)
axs[1].plot(K_range_IO, clusters_IO, marker='o', linestyle='-', color='b', linewidth=2)
axs[1].set_xlabel('Number of Clusters (k)', fontsize=12)
axs[1].set_ylabel('Within-Cluster Sum of Squares (WCSS)', fontsize=12)
axs[1].set_title('Elbow Method for merged_tourismIO', fontsize=14)
axs[1].grid(True, linestyle='--', alpha=0.7)

# Adjust layout and spacing
plt.tight_layout()

# Display the figure
plt.show()

from sklearn.metrics import silhouette_score

X = pca_df[["PC1", "PC2"]]

# Try different k-values
k_values = range(2, 11)
silhouette_scores = []

for k in k_values:
    kmeans = KMeans(n_clusters=k, random_state=42, n_init=10)
    kmeans.fit(X)
    labels = kmeans.labels_
    silhouette_avg = silhouette_score(X, labels)
    silhouette_scores.append(silhouette_avg)
    print(f"For k={k}, the silhouette score is {silhouette_avg:.3f}")

optimal_k = k_values[np.argmax(silhouette_scores)]

For k=2, the silhouette score is 0.567
For k=3, the silhouette score is 0.488
For k=4, the silhouette score is 0.541
For k=5, the silhouette score is 0.517
For k=6, the silhouette score is 0.404
For k=7, the silhouette score is 0.428
For k=8, the silhouette score is 0.461
For k=9, the silhouette score is 0.456
For k=10, the silhouette score is 0.463

X_IO = IO[["PC1_IO", "PC2_IO"]]

# Try different k-values
k_values_IO = range(2, 11)
silhouette_scores_IO = []

for k in k_values_IO:
    kmeans_IO = KMeans(n_clusters=k, random_state=42, n_init=10)
    kmeans_IO.fit(X_IO) # Changed from X to X_IO
    labels_IO = kmeans_IO.labels_ # Changed from labels to labels_
    silhouette_avg_IO = silhouette_score(X_IO, labels_IO)
    silhouette_scores_IO.append(silhouette_avg_IO)
    print(f"For k={k}, the silhouette score is {silhouette_avg_IO:.3f}")

optimal_k_IO = k_values_IO[np.argmax(silhouette_scores_IO)]

For k=2, the silhouette score is 0.572
For k=3, the silhouette score is 0.522
For k=4, the silhouette score is 0.526
For k=5, the silhouette score is 0.502
For k=6, the silhouette score is 0.380
For k=7, the silhouette score is 0.482
For k=8, the silhouette score is 0.384
For k=9, the silhouette score is 0.390
For k=10, the silhouette score is 0.421

# Create a figure with 2 subplots (1 row, 2 columns)
fig, axs = plt.subplots(1, 2, figsize=(18, 6))  # Adjust figsize as needed to accommodate 2 plots

# Plot silhouette score on the first subplot (axs[0])
axs[0].plot(k_values, silhouette_scores, marker='o', linestyle='-', color='navy', linewidth=2, label="Silhouette Score")
axs[0].axvline(optimal_k, color='crimson', linestyle='--', linewidth=2, label=f'Optimal k={optimal_k}')
axs[0].set_xlabel('Number of Clusters (k)', fontsize=12)
axs[0].set_ylabel('Silhouette Score', fontsize=12)
axs[0].set_title('Silhouette Score for merged_tourism', fontsize=14, fontweight='bold')
axs[0].legend(fontsize=12)
axs[0].grid(True, linestyle='--', alpha=0.7)
axs[0].set_xticks(k_values)
axs[0].tick_params(axis='x', labelsize=10)
axs[0].tick_params(axis='y', labelsize=10)

# Plot for silhouette score for merged_tourismIO
axs[1].plot(k_values_IO, silhouette_scores_IO, marker='o', linestyle='-', color='navy', linewidth=2, label="Silhouette Score")
axs[1].axvline(optimal_k_IO, color='crimson', linestyle='--', linewidth=2, label=f'Optimal k={optimal_k_IO}')
axs[1].set_xlabel('Number of Clusters (k)', fontsize=12)
axs[1].set_ylabel('Silhouette Score', fontsize=12)
axs[1].set_title('Silhouette Score for merged_tourismIO', fontsize=14, fontweight='bold')
axs[1].legend(fontsize=12)
axs[1].grid(True, linestyle='--', alpha=0.7)
axs[1].set_xticks(k_values_IO)
axs[1].tick_params(axis='x', labelsize=10)
axs[1].tick_params(axis='y', labelsize=10)

plt.tight_layout()  # Adjust subplot parameters for a tight layout
plt.show()  # Display the figure with all subplots

# Optimal K for merged_tourism
k_optimal = 4

# Fit K-Means
kmeans = KMeans(n_clusters=k_optimal, random_state=42, n_init=10)
pca_df["Cluster"] = kmeans.fit_predict(X)

# Optimial K for merged_tourismIO
k_optimal_IO = 4

# Fit K-Means
kmeans_IO = KMeans(n_clusters=k_optimal_IO, random_state=42, n_init=10)
IO["Cluster"] = kmeans_IO.fit_predict(X_IO)

import matplotlib.pyplot as plt
import seaborn as sns

# Create a figure with 2 subplots (1 row, 2 columns)
fig, axs = plt.subplots(1, 2, figsize=(16, 6))  # Adjust figsize as needed

# Scatter plot of clusters for merged_tourism (Plot 1)
# Before plotting, ensure the 'Cluster' column is in pca_df
pca_df["Cluster"] = kmeans.labels_  # Add this line to assign cluster labels
sns.scatterplot(x="PC1", y="PC2", hue="Cluster", data=pca_df, palette="viridis", s=50, ax=axs[0])
axs[0].scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:, 1], c="red", marker="X", s=200, label="Centroids")
axs[0].legend()
axs[0].set_title(f"K-Means Clustering (k={k_optimal}) for merged_tourism")

# Scatter plot of clusters for merged_tourismIO (Plot 2)
sns.scatterplot(x="PC1_IO", y="PC2_IO", hue="Cluster", data=IO, palette="magma", s=50, ax=axs[1])
axs[1].scatter(kmeans_IO.cluster_centers_[:, 0], kmeans_IO.cluster_centers_[:, 1], c="red", marker="X", s=200, label="Centroids")
axs[1].legend()
axs[1].set_title(f"K-Means Clustering (k={k_optimal_IO}) for merged_tourismIO")

plt.tight_layout()  # Adjust layout for better spacing
plt.show()

from scipy.cluster.hierarchy import linkage, dendrogram
import matplotlib.pyplot as plt


# Filter data for COVID-19 years (2020-2021)
covid_years_df = pca_df[pca_df['Year'].isin([2020, 2021])]

# Agglomerative Clustering for COVID data
linkage_matrix_covid = linkage(covid_years_df[['PC1', 'PC2']], method='ward') 

plt.figure(figsize=(12, 6))
dendrogram(linkage_matrix_covid, labels=covid_years_df['Country'].tolist(), orientation='top', leaf_font_size=8) 
plt.title('Hierarchical Clustering Dendrogram (COVID)')
plt.xlabel('Countries')
plt.ylabel('Distance')
plt.tight_layout()
plt.show()

# Dendogram All Merged_tourism 

linkage_matrix = linkage(pca_df[['PC1', 'PC2']], method='ward')

plt.figure(figsize=(12, 6))

dendrogram(linkage_matrix, labels=pca_df['Country'].tolist(), orientation='top', leaf_font_size=8)
plt.title('Hierarchical Clustering Dendrogram (All Merged_tourism)')
plt.xlabel('Countries')
plt.ylabel('Distance')

plt.tight_layout()
plt.show()

# Agglomerative Clustering for merged_tourismIO

# Filter data for COVID-19 years (2020-2021)
covid_years_df = IO[IO['Year'].isin([2020, 2021])]

# Agglomerative Clustering for COVID data
linkage_matrix_covid = linkage(covid_years_df[['PC1_IO', 'PC2_IO']], method='ward') 

plt.figure(figsize=(12, 6))
dendrogram(linkage_matrix_covid, labels=covid_years_df['Country'].tolist(), orientation='top', leaf_font_size=8) 
plt.title('Hierarchical Clustering Dendrogram (COVID)')
plt.xlabel('Countries')
plt.ylabel('Distance')
plt.tight_layout()
plt.show()

# Dendogram All Merged_tourismIO

linkage_matrix = linkage(IO[['PC1_IO', 'PC2_IO']], method='ward')

plt.figure(figsize=(12, 6))

dendrogram(linkage_matrix, labels=IO['Country'].tolist(), orientation='top', leaf_font_size=8)
plt.title('Hierarchical Clustering Dendrogram (All Merged_tourismIO)')
plt.xlabel('Countries')
plt.ylabel('Distance')

plt.tight_layout()
plt.show()

	Entity	Year	Share GDP	Inbound Arrivals	Business-Personal
0	Albania	2008	2.757073	1247000.0	0.058955
1	Albania	2009	2.668689	1711000.0	0.028825
2	Albania	2010	2.812342	2191000.0	0.019831
3	Albania	2011	2.534773	2469000.0	0.014884
4	Albania	2012	2.356966	3156000.0	0.012097
...	...	...	...	...	...
638	Uruguay	2017	8.500000	3674000.0	0.065889
639	Uruguay	2018	8.100000	3469000.0	0.068419
640	Uruguay	2019	7.000000	3059000.0	0.073641
641	Vietnam	2013	6.060000	7201000.0	0.200952
642	Vietnam	2014	6.140000	7488000.0	0.201770

	Entity	Year	Share GDP	Inbound Arrivals	Business-Personal	Domestic tourists	Foreign guests
0	Albania	2008	2.757073	1247000.0	0.058955	156000.0	57000.0
1	Albania	2009	2.668689	1711000.0	0.028825	171000.0	65000.0
2	Albania	2010	2.812342	2191000.0	0.019831	181000.0	74000.0
3	Albania	2011	2.534773	2469000.0	0.014884	182000.0	136000.0
4	Albania	2012	2.356966	3156000.0	0.012097	101000.0	149000.0
...	...	...	...	...	...	...	...
229	United Kingdom	2015	3.937450	35149000.0	0.353145	41820000.0	18937000.0
230	United Kingdom	2016	3.446720	37358000.0	0.344177	42490000.0	19792000.0
231	United Kingdom	2017	3.423722	39543000.0	0.303589	42510000.0	21833000.0
232	United Kingdom	2018	3.238154	38664000.0	0.302712	42040000.0	21704000.0
233	United Kingdom	2019	3.649353	39418000.0	0.293434	43490000.0	21697000.0

	Entity	Year	Share GDP	Inbound Arrivals	Business-Personal	Inbound-Outbound
0	Argentina	2016	1.917674	6668000.0	0.259773	0.640108
1	Argentina	2017	1.986258	6711000.0	0.298891	0.549496
2	Argentina	2018	1.839787	6942000.0	0.328241	0.623720
3	Argentina	2019	1.773273	7399000.0	0.252144	0.811828
4	Argentina	2020	0.883223	2090000.0	0.132143	0.735656
...	...	...	...	...	...	...
366	Uruguay	2015	6.500000	2773000.0	0.080539	1.250789
367	Uruguay	2016	7.300000	3037000.0	0.076592	1.770846
368	Uruguay	2017	8.500000	3674000.0	0.065889	2.053661
369	Uruguay	2018	8.100000	3469000.0	0.068419	1.781716
370	Uruguay	2019	7.000000	3059000.0	0.073641	1.391087

	Year	Share GDP	Inbound Arrivals	Business-Personal
count	643.000000	643.000000	6.430000e+02	643.000000
mean	2014.533437	4.427443	6.877075e+06	0.225166
std	3.753717	3.914314	1.252554e+07	0.322217
min	2008.000000	-0.100000	1.600000e+03	0.009049
25%	2011.000000	2.095105	6.600000e+05	0.061799
50%	2015.000000	3.238154	2.080000e+06	0.147796
75%	2018.000000	5.600000	7.408000e+06	0.265800
max	2021.000000	30.000000	9.091400e+07	4.076923

	Year	Share GDP	Inbound Arrivals	Business-Personal	Domestic tourists	Foreign guests
count	234.000000	234.000000	2.340000e+02	234.000000	2.340000e+02	2.340000e+02
mean	2014.769231	3.704931	1.184421e+07	0.174563	1.672148e+07	7.719913e+06
std	3.864407	2.033248	1.582227e+07	0.168655	2.300323e+07	1.216569e+07
min	2008.000000	0.690216	1.703000e+05	0.010295	4.800000e+04	3.400000e+04
25%	2011.000000	2.343264	2.400000e+06	0.048230	6.485000e+05	1.209750e+06
50%	2015.000000	3.227866	4.865000e+06	0.129556	4.710500e+06	3.227500e+06
75%	2018.000000	4.781707	1.338450e+07	0.287611	3.502925e+07	7.328000e+06
max	2021.000000	9.154534	9.091400e+07	1.699774	1.078640e+08	6.125000e+07

Project 4 Clustering: Tourism Through a Cluster Lens - Uncovering Global Patterns in GDP Contribution¶

Emily Santacruz Garcia¶

Introduction to the Problem¶

What is Clustering, and How Does it Work?¶

Introduce the Data¶

Merged Tourism Dataset (2008-2021):¶

Countries Under Review for This Dataset (Merged Tourism):¶

Merged Tourism: Local vs. Foreign Dataset (2008-2021):¶

Countries Under Review for This Dataset (Merged Tourism: Local vs. Foreign):¶

Merged Tourism: Inbound vs. Outbound Dataset (2008-2021):¶

Countries Under Review for This Dataset (Merged Tourism: Inbound vs. Outbound):¶

Data Understanding/Visualization¶

Pre-processing the Data¶

Conducting PCA on Datasets¶

'Merged_tourism' Dataset:¶

'Merged_tourismLF' Dataset:¶

'Merged_tourismIO' Dataset:¶

PCA Projection of the Datasets¶

Modeling (Clustering)¶

K-Means Clustering¶

Elbow Method to Determine Most Optimal K¶

'Merged_tourism' Dataset¶

'Merged_tourismIO' Dataset¶

Silhouette Score Method to Determine Most Optimal K¶

'Merged_tourism' Dataset¶

'Merged_tourismIO' Dataset¶

Conducting K-means Clustering¶

'Merged_tourism' Dataset¶

'Merged_tourismIO' Dataset¶

Hierarchical Clustering¶

'Merged_tourism' Dataset¶

'Merged_tourismIO' Dataset¶

Storytelling (Clustering Analysis)¶

Impact Section¶

References¶

Code¶

	Year	Share GDP	Inbound Arrivals	Business-Personal	Inbound-Outbound
count	371.000000	371.000000	3.710000e+02	371.000000	371.000000
mean	2014.563342	4.060410	8.544783e+06	0.179315	1.763662
std	3.835560	2.678303	1.457206e+07	0.141057	1.552501
min	2008.000000	-0.100000	2.400000e+04	0.018334	0.259563
25%	2011.000000	2.089153	1.340000e+06	0.073249	0.732487
50%	2015.000000	3.100000	2.927000e+06	0.145894	1.259950
75%	2018.000000	5.800000	9.315000e+06	0.240282	2.092894
max	2021.000000	13.500000	9.091400e+07	0.757194	15.484848

	PC1	PC2	Country	Year
0	0.058264	-0.259164	Albania	2008
1	-0.004448	-0.220394	Albania	2009
2	-0.059789	-0.201702	Albania	2010
3	-0.032198	-0.153167	Albania	2011
4	-0.026778	-0.084953	Albania	2012

	PC1_IO	PC2_IO	Country	Year
0	-1.078903	0.054046	Argentina	2016
1	-1.102097	-0.005412	Argentina	2017
2	-1.106968	0.067163	Argentina	2018
3	-1.038760	0.170539	Argentina	2019
4	-1.308802	0.371101	Argentina	2020